In a database grid, various events may be logged in order to better manage the system. For example, the system may be configured to log an event if a particular device in the network is overheating. When an event occurs on a node, an SNMP trap is typically sent from the node to an SNMP management node. The SNMP manager can then determine that an event has occurred. Centralized event management, in which only the management node has knowledge of events on other nodes, introduces delay and inefficiency in the system. In addition, having a single management node creates a single point of failure in the system. An improved method of event management is needed.